Turtlebot3 project presentation --autonomy

cu icar

Autononomy

Dr. Krovi curse, presentation

turtlebot3:https://www.turtlebot.com/

cost map:http://wiki.ros.org/costmap_2d

exploring turtelbot 3 and ros,

Ubuntu 1.604, install ros kinetic ,assemble turtle bot and install ros on tutlebot3

SLAM  simulations, g_mapping in gazebo eniroment, visualization in Rviz, RQT Graph.

mover package to move  turtle bot 3 in a straight line, simulated in gazebo

slam:tutlebot3_wold.launch, g-mapping metho

raspberry camera

line track:

chanllenges faced: light reflecting off the line, cause the camera to miss the line

Blob tracking:

detect the R/B color of RGB intensity and YUV threshold, set a bouding box around the detected color; dtermin the centroid

demo:mira 

line following, blob tracking ,leg detection obstacle detection.

open ground: light affects the color.

leg detection: use people_detection pkg from fetch robotics to determing the shape of leg by 2d lidar scans.

carrot controller: x,y detect leg

Terrain Obstacle Detection and Analysis using LIDARhttps://pdfs.semanticscholar.org/a795/2062e73af37a9cdb21e6581e8f9a2ab786ee.pdf

Obstacle avoiding: get the map of the envrioment by SLAM

motion planner: A*

code:python

tune the rotot base footprint, inflation radius and  the cost scaling factor in the navigation stack.

chanllege: localizaing,timesysnc between ros master and slave,

practise in gazebo, then integrate and transfer to turtlebot 3.

learned: ros, open source, python

Group 2

blob tracking:20 cm dsitance, by color  and centroid .

HSV, use one single value to detect the color.  mean value from the image dataset, particle filter works.

 use color and shape.

leg detect package, co-ordiantes of  the human wiht respect to hte odom frame.

http://wiki.ros.org/leg_detector

pid method was used to control the steer of the turtlebot 3.

chanllege: multi-person.

nav2d: error with gazebo.

amcl package:

 

Group3